Should Evidence be Requested for All New Drugs, or Just Some?

Australia requires economic submissions from manufacturers relating to all new drugs to be used outside public hospitals, as does Ontario (3;7;8,16). In contrast, The Netherlands only requires submissions for drugs outside the reference price system. Both Portugal and the United Kingdom are also quite selective regarding which drugs they subject to detailed appraisal.

One important factor is the availability of financial and human resources to undertake economic evaluations and to assess their quality. Also, the use of economic evaluation for some or all drugs needs to be viewed in the context of other pharmaceutical policies operating in the country concerned, for example, promotion of generic substitution, reference pricing, prescribing audits and so on.

From a Hungarian perspective it is important to define the term “new drug” from a practical point of view. Drugs fall into the “new drug” category when they first require coverage from the National Health Insurance Fund, or when requests are made to change the indications, the population covered, or the percentage of the subsidy. This is irrespective of how old or new the particular drug is. In Hungary, the number of available drugs at least partly covered by the health insurance increased greatly over the past 10 to 15 years. A sufficient evaluation capacity is needed for economic evaluation, so it is important to determine (realistically) what capacity would be considered as “sufficient” in Hungary for undertaking drug evaluations.

Taking the example of the National Institute of Clinical Excellence (NICE) in the United Kingdom, fifty appraisals were published in the first 3 years, in which seventy-four different drugs, ninety-four medical devices, fifteen clinical procedures, and forty different clinical conditions were evaluated. According to Buxton (2), “this is an impressive start but merely a scratch on the surface.” Buxton also stated that “NICE estimates its guidance to have the potential to increase costs to the NHS (in England and Wales) by approximately 575 million English Pounds.” Questions to be answered and lessons to be learned from the example of NICE are as follows: what capacity for appraisal seems to be realistic in the coming years in Hungary and what should be done about implementation of the results of health economic studies?

How Should Drugs be Prioritized for Assessment?

In cases where only certain drugs are appraised from an economic perspective, decisions are required on which to appraise (12;15). The most obvious criteria are: (i) drugs that tackle important health problems; and (ii) drugs that are likely to have a large budgetary impact. NICE considers appraising technologies that are likely to have a “major impact on the NHS.”

In Hungary, the intention is to use criteria similar to those applied by NICE. However, these criteria are very difficult to quantify and use in practice, due to the fact that no valid information source is available in Hungary relating to the morbidity and mortality pattern of the population. More evaluations and better databases (disease registers have just been created) need to be in place to identify a particular health problem, and the extent of the burden it creates. The “National Public Health Programme 2001–2010” does not help too much; nineteen priority categories are listed, but the public health items and the expected benefit are not specified in detail (17). The additional factor in the Hungarian situation is the very limited capacity for conducting health economics analyses. Therefore, drugs should be assessed which satisfy the following criteria:

Budget Impact Threshold for Economic Evaluation

If the estimated budgetary impact of the particular drug is greater than the 1 percent or 2 percent of the total drug budget in the coming 2 to 5 years, economic analysis might be initiated. (Similarly, in Australia if the estimated budgetary impact is bigger than Aus $2 million, a special committee decision is needed in addition to the standard procedure).

As Table 1 shows, a relatively small number of drugs with health insurance subsidies represent a significant percentage of the total budget for drug subsidies. Economic analysis might be particularly beneficial in this group of agents.

Analyzing the agents with the highest health insurance subsidies shows that even the agent in 25th place has a market share of 1 percent. The Top-10 agents have a total market share of 20.78 percent, while the Top-25 agents have 38.52 percent. These data suggest that evaluation of 25 agents would cover a substantial proportion of the total health insurance subsidies for drugs.

Should Data From Other Countries be Accepted? If so Which?

It is generally considered that clinical data for health technologies are transferable from location to location, although in all countries we encounter the problem that data from Phase III clinical trials may not reflect the effectiveness of the drug in regular clinical practice. Setting this problem aside, clinical data from another country are usually acceptable, unless the patients enrolled in the trial, or the regimens compared, are different from those that would apply in the country of interest. Data on practice patterns, and hence resource use, are likely to be less transferable and some additional data collection in the country of interest is likely to be required, if only to check that data from another country are relevant. It is almost certain that unit cost or price data from another country will not apply and these should be collected for the country of interest (14). The case for new data collection of health state preference values (or utilities) is less clear. NICE requires that the preferences should be relevant for the population of England and Wales, but this does not necessarily mean that preference values from elsewhere are not relevant. The issue of transferability of economic data is crucial for small middle income countries where the capacity for undertaking health economic analysis is limited. This has been discussed by Rutten and Gulácsi (14), who highlighted the most important items that need to be taken into consideration.

Should an Independent Assessment be Made of Evidence, Including Company Submissions?

In all countries operating a fourth hurdle, assessments are made of the relevant evidence on cost-effectiveness as part of the decision-making process. An important component of the evidence is contained in submissions from manufacturers and sponsors of the health technologies concerned. Whilst these often contain high quality analyses from consultants or academic researchers funded by the company, submissions should be viewed as advocacy for the product (i.e., attempts to show the product in the best possible light). The lengths to which various jurisdictions go to assess the evidence in company submissions varies greatly. In some places, especially the smaller countries, the assessments are undertaken in-house, within the Ministry or drug reimbursement agency. In other jurisdictions the in-house staff are supported by a committee of academic economists and health service researchers who also assess the submissions (e.g., in Australia).

At the other end of the spectrum, organizations like NICE in the United Kingdom commission their own independent assessment of the evidence from academic research centres. These technology assessment reviews include not only an assessment of the company submissions, but also meta-analyses of the clinical literature and additional economic models.

Should a Two-Stage Appraisal Process be Considered?

The main argument in favor of a two-stage process is that, at the time of launch of a new drug, very little is known about its long-term benefits when used in regular clinical practice. Therefore, the initial economic assessment has to be based on efficacy data from Phase III clinical studies, perhaps augmented by modelling. In some countries the existing arrangements already provide for the possibility of gathering additional data and/or a review of guidance at a later date (e.g., 3 years). On an intellectual level this makes a lot of sense.The main difficulties are practical ones. For example, if the new drug is approved for reimbursement, will the sponsoring company be willing to consider further randomized controlled trials? Who will pay for the studies, including the study drug, post-reimbursement? How feasible would it be to remove a drug from the reimbursement list, or at least restrict its use, should the longer-term data be unimpressive? On the other hand, the longer-term cost-effectiveness results may be more impressive if clinicians learn to target the drug effectively, or obtain satisfactory therapeutic results by using a lower dose than was used in the clinical trials. In principle, a risk-sharing deal is one way around some of the difficulties outlined above. An example of one such deal is the agreement in the UK between the government and the manufacturers of beta interferon for multiple sclerosis. In this case NICE was skeptical about some of the long-term outcomes, in particular the clinical outcomes for patients who have to discontinue therapy. After much argument, the government decided to allow treatment for certain categories of patients, at the drug price requested by the manufacturers. However, it intends to monitor the long-term progress of patients and may ask the companies to return some of the income they receive from sales of the drugs if the results are less impressive than claimed (1). This is a very promising drug coverage option in Hungary, in cases where the patient population is small, the clinical outcome is well defined and the treatment process itself is well controlled. These criteria apply for those drugs that are covered by a special budget, in accredited hospitals, administered by named and accredited provider organizations and doctors, using defined clinical protocols on named patients.